如何将numpy数组列表转为Pandas DataFrame？解决二维输入报错问题

The error you're seeing happens because each element in your list_arrays is a 2D NumPy array (shape (1,9)), and pd.DataFrame() expects either a 2D array directly, or a list of 1D sequences (like 1D arrays or lists) to represent rows. When you pass a list of 2D arrays, pandas can't interpret them as valid row inputs, hence the ValueError: Must pass 2-d input.

Here are a few straightforward fixes:

1. Extract the 1D row from each 2D array #

Since each array in your list is a single row wrapped in an extra dimension, you can simply index the first (and only) row of each array to get a 1D array, then pass that list to pd.DataFrame:

import pandas as pd
import numpy as np

list_arrays = [np.array([[0, 0, 0, 1, 0, 0, 0, 0, 0]], dtype='uint8'), 
               np.array([[0, 0, 3, 2, 0, 0, 0, 0, 0]], dtype='uint8')]

# Extract the first row of each 2D array
d = pd.DataFrame([arr[0] for arr in list_arrays])

Alternatively, use flatten() to remove the extra dimension:

d = pd.DataFrame([arr.flatten() for arr in list_arrays])

2. Concatenate all arrays into a single 2D array first #

Another approach is to combine all your 2D arrays into one large 2D array using np.concatenate, then feed that directly to pd.DataFrame:

# Concatenate along the row axis (axis=0)
combined_array = np.concatenate(list_arrays, axis=0)
d = pd.DataFrame(combined_array)

Both methods will produce a DataFrame with each original array as a row, resulting in a shape of (2,9) (2 rows, 9 columns).

内容的提问来源于stack exchange，提问作者Marcos Santana